Oracle Restart环境下的数据库开机自动启动

最新推荐文章于 2024-01-04 14:03:16 发布

dingdingfish

最新推荐文章于 2024-01-04 14:03:16 发布

阅读量644

点赞数

分类专栏： Oracle 19c 文章标签： oracle 12c

本文链接：https://blog.csdn.net/stevensxiao/article/details/100688841

版权

Oracle 19c 专栏收录该内容

184 篇文章

订阅专栏

Linux 7环境，假设已安装GI，已基于ASM建立数据库。我们来看一下数据库是如何自动启动的。

Linux 7使用systemd管理服务，systemd 是系统启动后的第一个进程：

$ ps --pid 1
  PID TTY          TIME CMD
    1 ?        00:00:01 systemd

Oracle Restart的概念是操作系统包含OHAS，OHAS包含其他资源，如ASM，数据库等。
在服务中我们找到了ohas（–type表示只列出服务，–all表示不论状态）：

$ systemctl list-units --type service --all|grep ohas
  UNIT                                                  LOAD      ACTIVE   SUB     DESCRIPTION
  ohasd.service                                         loaded    active   exited  LSB: Start and Stop Oracle High Availability Service
  oracle-ohasd.service                                  loaded    active   running Oracle High Availability Services

启动脚本中链接到了ohas脚本：

$ ll /etc/rc3.d/S96ohasd
lrwxrwxrwx. 1 root root 17 Sep  9 01:37 /etc/rc3.d/S96ohasd -> /etc/init.d/ohasd

这两个服务状态都是active的，但子状态有一个是exited:

$ systemctl is-active oracle-ohasd.service
active
$ systemctl is-active ohasd.service
active

看一下此两服务的状态：

$ systemctl status ohasd.service
● ohasd.service - LSB: Start and Stop Oracle High Availability Service
   Loaded: loaded (/etc/rc.d/init.d/ohasd; bad; vendor preset: disabled)
   Active: active (exited) since Tue 2019-09-10 00:41:44 UTC; 58min ago
     Docs: man:systemd-sysv-generator(8)
  Process: 1265 ExecStart=/etc/rc.d/init.d/ohasd start (code=exited, status=0/SUCCESS)
 
$ systemctl status oracle-ohasd.service
● oracle-ohasd.service - Oracle High Availability Services
   Loaded: loaded (/etc/systemd/system/oracle-ohasd.service; enabled; vendor preset: disabled)
   Active: active (running) since Tue 2019-09-10 00:41:43 UTC; 58min ago
 Main PID: 1262 (init.ohasd)
   CGroup: /system.slice/oracle-ohasd.service
           ├─1262 /bin/sh /etc/init.d/init.ohasd run >/dev/null 2>&1 </dev/null
           ├─1334 /u01/app/19.0.0/grid/bin/ohasd.bin reboot
           ├─2540 /u01/app/19.0.0/grid/bin/oraagent.bin
           ├─2646 /u01/app/19.0.0/grid/bin/tnslsnr LISTENER -no_crs_notify -inherit
           ├─2666 /u01/app/19.0.0/grid/bin/evmd.bin
           ├─3022 /u01/app/19.0.0/grid/bin/evmlogger.bin -o /u01/app/19.0.0/grid/log/[HOSTNAME]/evmd/evmlogger.info -l /u01/app/19.0.0/grid/log/[HOSTNAME]/...
           ├─3083 /u01/app/19.0.0/grid/bin/cssdagent
           ├─3199 /u01/app/19.0.0/grid/bin/ocssd.bin
           ├─4599 asm_pmon_+ASM
           ├─4601 asm_clmn_+ASM
           ├─4603 asm_psp0_+ASM
           ├─4606 asm_vktm_+ASM
           ├─4610 asm_gen0_+ASM
           ├─4612 asm_mman_+ASM
           ├─4616 asm_gen1_+ASM
           ├─4619 asm_diag_+ASM
           ├─4621 asm_pman_+ASM
           ├─4623 asm_dia0_+ASM
           ├─4625 asm_dbw0_+ASM
           ├─4627 asm_lgwr_+ASM
           ├─4629 asm_ckpt_+ASM
           ├─4631 asm_smon_+ASM
           ├─4633 asm_lreg_+ASM
           ├─4635 asm_pxmn_+ASM
           ├─4637 asm_rbal_+ASM
           ├─4639 asm_gmon_+ASM
           ├─4641 asm_mmon_+ASM
           ├─4643 asm_mmnl_+ASM
           ├─4705 oracle+ASM (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))
           └─4706 oracle+ASM (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))

ohasd.service负责启动ohas服务，已经退出。
oracle-ohasd.service正在运行，是实际的服务守护进程。

看一下这两个服务执行的命令：

$ systemctl show oracle-ohasd.service|grep Exec
ExecMainStartTimestamp=Tue 2019-09-10 00:41:43 UTC
ExecMainStartTimestampMonotonic=11457980
ExecMainExitTimestampMonotonic=0
ExecMainPID=1262
ExecMainCode=0
ExecMainStatus=0
ExecStart={ path=/etc/init.d/init.ohasd ; argv[]=/etc/init.d/init.ohasd run >/dev/null 2>&1 </dev/null ; ignore_errors=no ; start_time=[Tue 2019-09-10 00:41:43 UTC] ; stop_time=[n/a] ; pid=1262 ; code=(null) ; status=0/0 }
ExecStop={ path=/etc/init.d/init.ohasd ; argv[]=/etc/init.d/init.ohasd stop >/dev/null 2>&1 </dev/null ; ignore_errors=no ; start_time=[n/a] ; stop_time=[n/a] ; pid=0 ; code=(null) ; status=0/0 }

$ systemctl show ohasd.service|grep Exec
ExecMainStartTimestampMonotonic=0
ExecMainExitTimestampMonotonic=0
ExecMainPID=0
ExecMainCode=0
ExecMainStatus=0
ExecStart={ path=/etc/rc.d/init.d/ohasd ; argv[]=/etc/rc.d/init.d/ohasd start ; ignore_errors=no ; start_time=[Tue 2019-09-10 00:41:43 UTC] ; stop_time=[Tue 2019-09-10 00:41:44 UTC] ; pid=1265 ; code=exited ; status=0 }
ExecStop={ path=/etc/rc.d/init.d/ohasd ; argv[]=/etc/rc.d/init.d/ohasd stop ; ignore_errors=no ; start_time=[n/a] ; stop_time=[n/a] ; pid=0 ; code=(null) ; status=0/0 }

查看一下上面输出中的ExecMainPID:

$ ps p 1262
  PID TTY      STAT   TIME COMMAND
 1262 ?        Ss     0:00 /bin/sh /etc/init.d/init.ohasd run >/dev/null 2>&1 </dev/null

我们注意到，有两个关键的脚本，ohasd和init.ohasd，在/etc/init.d和/etc/rc.d/init.d下都有，互为hard link，因为inode相同：

$ ls -l /etc/init.d/*ohas*
-rwxr-x---. 1 root root 15119 Sep  9 01:37 /etc/init.d/init.ohasd
-rwxr-x---. 1 root root  7791 Sep  9 01:37 /etc/init.d/ohasd

$ ls -l /etc/rc.d/init.d/*ohas*
-rwxr-x---. 1 root root 15119 Sep  9 01:37 /etc/rc.d/init.d/init.ohasd
-rwxr-x---. 1 root root  7791 Sep  9 01:37 /etc/rc.d/init.d/ohasd

$ ls -i /etc/init.d/ohasd /etc/rc.d/init.d/ohasd
83447865 /etc/init.d/ohasd  83447865 /etc/rc.d/init.d/ohasd

ohasd虽然有300多行，但核心就一句话，启动ohas服务：

$ grep "crsctl start" ohasd
      log_console "Fix the problem and issue command 'crsctl start has' as $HAS_USER user to start Oracle Grid Infrastructure."
        my_crsctl start has -nowait

再来看iint.ohasd，有500多行。他的一个作用是监控ohas的状态，并在错误时重启。

看一下关于has的进程，init.ohasd是root运行的，而ohasd.bin是GI用户运行的，可能是实际的has服务进程：

$ ps -ef|grep ohas
root      1262     1  0 00:41 ?        00:00:00 /bin/sh /etc/init.d/init.ohasd run >/dev/null 2>&1 </dev/null
grid      1334     1  0 00:41 ?        00:00:40 /u01/app/19.0.0/grid/bin/ohasd.bin reboot
grid     10586  4880  0 02:25 pts/0    00:00:00 grep --color=auto ohas

最后再来说一下dbstart和dbshut，可用来启停数据库：

$ which dbstart dbshut
/u01/app/19.0.0/grid/bin/dbstart
/u01/app/19.0.0/grid/bin/dbshut

其它

参考

https://oracle-base.com/articles/linux/automating-database-startup-and-shutdown-on-linux
https://www.digitalocean.com/community/tutorials/how-to-use-systemctl-to-manage-systemd-services-and-units
https://unix.stackexchange.com/questions/396630/the-proper-way-to-test-if-a-service-is-running-in-a-script
https://orainternals.wordpress.com/2013/06/05/clusterware-startup/
https://zhuanlan.zhihu.com/p/54221584